You are on page 1of 12

Food Hydrocolloids 19 (2005) 209–220

www.elsevier.com/locate/foodhyd

Influence of environmental stresses on stability of O/W emulsions


containing droplets stabilized by multilayered membranes produced
by a layer-by-layer electrostatic deposition technique
Tomoko Aoki, Eric A. Decker, D. Julian McClements*
Biopolymers and Colloids Research Laboratory, Department of Food Science, University of Massachusetts, Amherst, MA 01003, USA
Received 26 March 2004; revised 19 May 2004; accepted 20 May 2004

Abstract
A three step process, based on electrostatic layer-by-layer deposition, was used to produce oil-in-water emulsions containing submicron
(d32 < 0:3 mm) oil droplets stabilized by sodium dodecyl sulfate (SDS) – chitosan– pectin membranes (100 mM acetic acid, pH 3.0).
First, primary emulsions (5 wt% corn oil, 5 mM SDS) containing anionic droplets (z < 260 mV) stabilized by SDS membranes were
prepared using a high-pressure value homogenizer. Second, secondary emulsions (1 wt% corn oil, 1 mM SDS, 0.024 wt% chitosan)
containing cationic droplets (z < þ59 mV) stabilized by SDS – chitosan membranes were formed by diluting the primary emulsion with
aqueous chitosan solution and agitating to disrupt any flocs. Third, tertiary emulsions (0.2 wt% corn oil, 0.2 mM SDS, 0.0048 wt% chitosan,
0.040 wt% pectin) containing anionic droplets (z < 214 mV) stabilized by SDS – chitosan –pectin membranes were formed by diluting the
secondary emulsion with aqueous pectin solution. The stability of primary, secondary and tertiary emulsions to pH, ionic strength, thermal
processing and freeze – thaw cycling were determined. The droplets in tertiary emulsions had good stability to droplet aggregation over a
wide range of pH values (pH 3.0– 8.0), NaCl concentrations (#500 mM NaCl), thermal treatments (30– 90 8C for 20 min) and freeze – thaw
cycling (2 20 8C for 22 h/30 8C for 2 h). The interfacial engineering technology utilized in this study could lead to the creation of food
emulsions with improved stability to environmental stresses.
q 2004 Elsevier Ltd. All rights reserved.
Keywords: Emulsion stability; Layer-by-layer deposition; SDS; Chitosan; Pectin

1. Introduction Krog & Sparso, 2004; Stauffer, 1999). Each type of


emulsifier varies in its effectiveness at producing small oil
Conventionally, oil-in-water emulsions are created by droplets during homogenization, and its ability to prevent
homogenizing oil and aqueous phases together in the droplet aggregation under different environmental stresses,
presence of one or more emulsifiers (Friberg, Larsson, & such as pH, ionic strength, heating, and freezing (Dickinson,
Sjoblom, 2004; McClements, 1999). An emulsifier is a 2003; Garti & Reichman, 1993; McClements, 1999).
surface-active molecule that adsorbs to the surface of the Food emulsifiers also differ in their cost, reliability, ease of
emulsion droplets to form a protective membrane that use, ingredient compatibility, ‘label friendliness’ and legal
prevents them from aggregating (flocculating and/or status (Krog & Sparso, 2004; Stauffer, 1999). For these
coalescing) with one another. In addition, an emulsifier reasons, there is no single emulsifier that is ideal for use in
reduces the oil –water interfacial tension, thereby facilitating every type of food product. Instead, the selection of the most
the disruption of emulsion droplets during homogenization appropriate emulsifier or combination of emulsifiers for a
(Walstra, 1993, 2003). A wide variety of different emulsifiers particular food product depends on the type and
are available for utilization in the food industry, including concentration of other ingredients that it contains, the way
proteins, polysaccharides, phospholipids and small that it was produced, and the environmental conditions that it
molecule surfactants (Charlambous & Doxastakis, 1989; experiences during its manufacture, storage and utilization.
There are limitations to the functional properties that
* Corresponding author. Tel.: þ 1-413-545-1019; fax: þ1-413-545-1262. can be achieved using existing food emulsifiers
E-mail address: mcclements@foodsci.umass.edu (D.J. McClements). and the conventional method of creating emulsions, which
0268-005X/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodhyd.2004.05.006
210 T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220

has led to research being carried out to find alternative NaCl or CaCl2), as well as exhibiting less lipid oxidation
methods of improving emulsion stability by developing (peroxide formation) (Ogawa et al., 2003a,b).
novel emulsifier-based strategies. One strategy has been to Nevertheless, the secondary emulsions were only stable
create covalent protein – polysaccharide complexes that at relatively low pH values (, pH 5) and required
have good surface activity and provide improved protection cryoprotectants to inhibit freeze –thaw instability, which
against environmental stresses (Dickinson, 2003; Schmitt, would limit their application in many industrial products.
Sanchez, Desobry-Banon, & Hardy, 1998). An alternative The objective of the present investigation was to determine
strategy is to create an interfacial membrane around oil whether we could improve the freeze – thaw and pH
droplets that consists of multiple layers of emulsifiers stability of lecithin– chitosan emulsions by incorporating
and/or biopolymers using a layer-by-layer deposition pectin as a third layer.
technique (Calvo, Remunan-Lopez, Vila-Jato, & Alonso,
1997; Caruso, 2001; Caruso & Schuler, 2000; Dickinson
& James, 2000; Faldt, Bergenstahl, & Claesson 1993; 2. Experimental
Magdassi, Bach, & Mumcuoglu 1997; Moreau, Kim,
Decker, & McClements 2003; Ogawa, Decker, & 2.1. Materials
McClements 2003a,b; Shi & Caruso, 2001). In this
approach, an ionic emulsifier that rapidly adsorbs to the Sodium dodecyl sulfate (SDS), chitosan (from crab
surface of lipid droplets during homogenization is used to shells, minimum 85% deacetylated), analytical grade
produce a primary emulsion containing small droplets, then sodium chloride, hydrochloric acid, sodium hydroxide,
an oppositely charged biopolymer is added to the system sodium azide, acetic acid and sodium acetate were obtained
that adsorbs to the droplet surfaces and produces from the Sigma Chemical Company (St Louis, MO).
secondary emulsions containing droplets coated with a Pectin (extracted from citrus) was purchased from Sigma-
two-layer interfacial membrane. This latter procedure can Aldrich Co. (St Louis, MO). As stated by the manufacturer,
be repeated to form oil droplets coated by interfacial the degree of esterification of the pectin was 59%. Corn oil
membranes containing three or more layers (Caruso, 2001; was purchased from a local supermarket and used without
Ogawa, Decker, & McClements 2004). Under certain further purification. Distilled and de-ionized water was used
circumstances, emulsions containing oil droplets for the preparation of all solutions.
surrounded by multilayered interfacial membranes have
been found to have better stability to environmental 2.2. Solution preparation
stresses than conventional oil-in-water emulsions with
single-layered interfacial membranes. An emulsifier solution was prepared by dispersing
The layer-by-layer deposition method therefore offers a 20 mM SDS in acetate buffer (100 mM acetic acid, adjusted
promising way to improve emulsion stability, but the to pH 3.0 with 1 M HCl). Chitosan and pectin solutions were
choice of an appropriate combination of emulsifier and prepared by dispersing 0.2 wt% chitosan and 0.4 wt% pectin
biopolymers is essential. In this study we examine the use in acetate buffer. Sodium azide stock of 0.2 wt% was
of an anionic surfactant (SDS), a cationic biopolymer prepared in acetate buffer. The pH of these solutions was
(chitosan), and an anionic biopolymer (pectin) to form the adjusted back to 3.0 using 1 M HCl, and then the solutions
interfacial membranes. SDS was chosen to represent were stirred for about 1 day to ensure complete dissolution
small-molecule anionic surfactants commonly used in of these materials.
foods, such as fatty acid salts, lecithin, SLS, DATEM,
and CITREM (Krog & Sparso, 2004). Chitosan (cationic) 2.3. Primary emulsion preparation
and pectin (anionic) were chosen because they are
examples of food-grade polysaccharides that have opposite A preliminary experiment was carried out to determine
electrical charges. Chitosan is positively charged and the optimum SDS concentration required to create the
water-soluble at low pH (pKa < 6.5), but loses its charge primary emulsions, i.e. the minimum amount of SDS
at higher pH and may precipitate from solution (Shahidi, required to produce emulsions that were stable to droplet
Arachchi, & Jeon, 1999; Skjak-Braek, Anthonsen, & aggregation. Primary emulsions were prepared by
Sandford, 1989). Pectin is negatively charged at relatively homogenizing 5 wt% corn oil with 95 wt% aqueous
high pH values (pKa < 4 – 5), and is water-soluble emulsifier solution (1 –10 mM SDS) in a high-speed blender
(Williams & Phillips, 2003). Previous studies have shown (M133/1281-0, Biospec Products, Inc., ESGC, Switzerland)
that secondary emulsions stabilized by lecithin– chitosan followed by sonication for 30 s at a frequency of 20 kHz,
membranes have better stability than primary emulsions amplitude of 40% and duty cycle of 0.5 s (Model 500,
stabilized by lecithin membranes to droplet aggregation Sonic Disembrator, Fisher Scientific, Pittsburgh, PA).
brought about by thermal processing (30 – 90 8C for The primary emulsions were then stored at room
30 min), freeze – thaw cycling (2 10 8C for 22 h/30 8C temperature for 24 h before being analyzed to determine
for 2 h) and exposure to high ionic strengths (, 500 mM the free SDS concentration and their stability to droplet
T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220 211

aggregation. The optimum surfactant concentration was 2.7. Particle size measurements
established to be 5 mM SDS (see Section 3.1), which was
used to prepare the primary emulsions in all subsequent Emulsions were diluted to a droplet concentration of
experiments. approximately 0.005 wt% oil using acetate buffer solution
to avoid multiple scattering effects. The particle size
distribution of the emulsions was measured using a laser
2.4. Secondary emulsion preparation
light scattering instrument (Mastersizer, Malvern
Instruments Ltd., Worcs., UK). This instrument measures
A primary emulsion was prepared by homogenizing the angular dependence of the intensity of laser light
5 wt% corn oil with 95 wt% aqueous emulsifier solution (l ¼ 632:8 nm) scattered by a dilute emulsion, and then
(5 mM SDS) in a high-speed blender (M133/1281-0, finds the particle size distribution that gives the best
Biospec Products, Inc., ESGC, Switzerland) followed by agreement between theoretical predictions and experimental
three passes at 5000 psi through a two-stage high-pressure measurements. A refractive index ratio of 1.08 was used in
valve homogenizer (LAB 1000, APV-Gaulin, Wilmington, the calculations of the particle size distribution. It should
MA). This emulsion was diluted with aqueous chitosan be noted that the theory used to calculate the particle size
solutions to form secondary emulsions with varying distribution assumes that the particles are spherical and
compositions: 1 wt% corn oil, 1 mM SDS, 100 mM acetic homogeneous, and therefore the data obtained on emulsions
acid, and 0– 0.040 wt% chitosan (pH 3.0). The secondary that contained flocs should be treated with caution because
emulsions were stored at room temperature for 24 h before they are non-spherical and non-homogenous. Particle size
being analyzed. measurements are reported as the volume-surface mean
diameter: d32 ¼ Sn3i =Sn2i ; where ni is the number of droplets
2.5. Tertiary emulsion preparation of diameter di : Mean particle diameters were calculated as
the average of measurements made on at least three
samples, with standard deviations being less than 10%.
A secondary emulsion was prepared as described in
Section 2.4, that had a final concentration of 1 wt% corn oil,
1 mM SDS, 0.024 wt% chitosan, 100 mM acetic acid 2.8. z-Potential measurements
(pH 3.0). This system was stirred for 1 h using a magnetic
stirrer at ambient temperature to thoroughly mix the primary Emulsions were diluted to a droplet concentration of
emulsion with the chitosan. As observed in previous studies, approximately 0.005 wt% oil using buffer solution to avoid
some droplet flocculation was observed in the secondary multiple scattering effects. Diluted emulsions were
emulsion prepared using this method, so the emulsions was injected directly into the measurement chamber of a
passed three times through a high-pressure value particle electrophoresis instrument (ZEM5003, Zetamaster,
homogenizer at a pressure of 5000 psi to disrupt these Malvern Instruments, Worcs., UK). The z-potential
flocs (Ogawa et al 2003a,b). Tertiary emulsions were was then determined by measuring the direction and
formed by diluting the secondary emulsion with aqueous velocity of droplet movement in a well-defined electric
pectin solution to produce a series of emulsions with field. The z-potential measurements are reported as the
different pectin concentrations: 0.2 wt% corn oil, 0.2 mM mean and standard deviation of two separate injections, with
SDS, 0.0048 wt% chitosan, 100 mM acetic acid and five readings made per injection.
0 –0.040 wt% pectin (pH 3.0). The tertiary emulsions were
stored at room temperature for 24 h before being analyzed. 2.9. Creaming stability measurements

Approximately 3.5 g samples of diluted emulsion


2.6. Free SDS measurements (0.005 wt% oil) were transferred into 1-cm path length
plastic spectrophotometer cuvettes, and then stored at room
A surfactant selective electrode (SSE) was used to temperature for 1 day. The turbidity (at 600 nm) of the
measure electro-motive force (EMF) of free SDS. The SSE emulsions was then measured using an UV – visible
consisted of a surfactant selective electrode, a double spectrophotometer. The light beam passed through the
junction reference electrode, and a pH meter (Thermo Orion, emulsions at a height that was 10 mm from the cuvette
Beverly, MA). The EMF of primary emulsions prepared bottom, i.e. about 30% of the emulsion’s height. The oil
using different initial surfactant concentrations was droplets in the emulsions moved upwards due to
measured by dipping the SSE electrode directly into the gravity, which led to the formation of a relatively clear
emulsions (1 –10 mM SDS). The free SDS concentration in droplet-depleted serum layer at the bottom of the cuvette.
the emulsions was then calculated from a standard curve An appreciable decrease in emulsion turbidity was therefore
prepared using known concentrations of SDS dissolved in an indication of the fact that the serum layer had risen to at
acetate buffer (100 mM, pH 3.0). least 30% of the emulsion’s height.
212 T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220

2.10. Emulsion environmental stresses

We compared the influence of various kinds of


environmental stresses on the z-potential, mean particle
diameter and creaming stability of primary, secondary and
tertiary emulsions.
Thermal processing. The primary and secondary emul-
sions were diluted with acetic acid buffer (pH 3.0) to adjust
the oil concentration to the same value as in the tertiary
emulsion (0.2 wt%). Emulsion samples (10 ml) were
transferred into glass test tubes, which were then incubated
in a water bath for 30 min at different temperatures ranging
from 30 to 90 8C. After incubation the emulsion samples
were immediately placed at room temperature, where they
were stored prior to analysis.
Freeze – thaw cycling stability. The primary and Fig. 1. Dependence of free SDS concentration on total SDS concentration in
primary emulsions (5 wt% corn oil, 100 mM acetic acid buffer, pH 3.0)
secondary emulsions were diluted with acetic acid buffer measured using a surfactant selective electrode.
(pH 3.0) to adjust the oil concentration to the same value as
in the tertiary emulsion (0.2 wt%). These emulsion samples
emulsion was studied by measuring the EMF with a
(10 ml) were transferred into cryogenic test tubes and were
surfactant selective electrode (SSE). A standard curve of
incubated in a 2 20 8C freezer for 22 h. After incubation the
EMF versus free SDS was obtained by measuring the EMF
emulsion samples were thawed by incubating them in a
of a series of solutions with different SDS concentrations
water bath at 30 8C for 2 h. This freeze –thaw cycle was
(100 mM acetic acid buffer, pH 3.0). The free SDS
repeated six times and its influence on emulsion properties
concentration in primary emulsions (5 wt% corn oil,
was measured after each cycle.
100 mM acetic acid buffer, pH 3.0) containing different
NaCl stability. Primary, secondary and tertiary emulsion
total SDS concentrations (1 – 10 mM) were then determined
samples were diluted with acetic acid buffer (pH 3.0) to
from EMF measurements using the standard curve (Fig. 1).
obtain the same final oil concentration (0.05 wt%),
Fig. 1 shows that around 2– 3 mM of SDS was adsorbed to
but different NaCl concentrations (0 – 500 mM).
the surfaces of the emulsion droplets, and above this value
The influence of NaCl concentration (0 – 500 mM) on the
the rest of the SDS was present in the continuous phase.
z-potential, mean particle diameter, and creaming stability
Ideally, we wanted to minimize the concentration of non-
of diluted primary, secondary and tertiary emulsions at pH
adsorbed SDS in the primary emulsions (to avoid its
3.0 was measured after 24 h.
interaction with chitosan), but we wanted to use sufficient
pH stability. Primary, secondary and tertiary emulsion
SDS to ensure that small droplets were formed that were
samples were diluted with acetate buffer (pH 3.0 –8.0) to
obtain the same final oil concentration (0.05 wt%), relatively stable to aggregation. We found that if the total
but different pH values. The influence of pH (3.0 – 8.0) on SDS concentration was too low (, 4 mM), then the droplets
the z-potential, mean particle diameter, and creaming formed were relatively large or were unstable to
stability of diluted primary, secondary and tertiary coalescence (i.e. oiling off was observed on the surface of
emulsions was measured after 1 day and 10 days storage. the emulsions). We therefore decided to use 5 mM SDS to
prepare the primary emulsions used in the subsequent
experiments.
3. Results and discussion
3.2. Influence of chitosan concentration on droplet
3.1. Influence of SDS concentration on droplet characteristics of secondary emulsions
characteristics of primary emulsions
The purpose of these experiments was to establish the
The purpose of these experiments was to determine the optimum chitosan concentration required to prepare stable
optimum SDS concentration required to prepare the primary secondary emulsions. The z-potential, mean particle
emulsions. If the SDS concentration is too low, it would not diameter and turbidity of secondary emulsions (1 wt%
be possible to form stable emulsions, but if the SDS corn oil, 1 mM SDS, 100 mM acetic acid buffer, pH 3.0)
concentration was too high there would be a lot of containing different chitosan concentrations (0 –0.040 wt%)
non-adsorbed surfactant present in the aqueous phase that were measured after 24 h storage at room temperature
could interact with the biopolymers subsequently used to (Fig. 2). In the absence of chitosan, the net charge on the
create the multiple-layered interfacial membranes. emulsion droplets was , 2 60 mV, because the SDS used
The adsorption of SDS to oil droplets in the primary to stabilize the droplets in the primary emulsion has
T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220 213

Fig. 2. (a) Dependence of particle electrical charge (z-potential) on chitosan concentration for secondary emulsions (1 wt% corn oil, 1 mM SDS, 100 mM
acetic acid, pH 3.0). (b) Dependence of mean particle diameter ðd32 Þ on chitosan concentration for secondary emulsions (1 wt% corn oil, 1 mM SDS, 100 mM
acetic acid, pH 3.0). The photographs at the top show micrographs of selected emulsions with different chitosan concentrations obtained by optical
microscopy. (c) Dependence of emulsion creaming stability on chitosan concentration for secondary emulsions (1 wt% corn oil, 1 mM SDS, 100 mM acetic
acid, pH 3.0). The creaming stability was determined as turbidity (at 600 nm) measured at 30% of emulsion height after 24 h storage: creaming instability is
indicated by a low turbidity. The photograph shown above highlights the instability of the emulsions to creaming at intermediate chitosan concentrations
(taken after sonication).
214 T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220

a negative charge at pH 3 (Fig. 2a). The net charge on the of the emulsions were measured (Fig. 2). Sonication caused
droplets switched from negative to positive when the a pronounced decrease in the mean particle diameter of all
chitosan concentration in the emulsions was increased, the emulsions (Fig. 2b), which suggests that it was
and eventually reached a relatively constant positive value capable of generating forces sufficiently large to either
(, þ 60 mV) when the chitosan concentration exceeded physically cleave chitosan molecules or to cause them to
about 0.020 wt% (Fig. 2a), presumably because the droplets become detached from all but one of the droplets.
become saturated with chitosan. This change in z-potential Despite the application of sonication, emulsions containing
suggests that cationic chitosan molecules adsorbed to the 0.012 – 0.016 wt% chitosan still exhibited extensive
surface of anionic SDS-coated emulsion droplets. aggregation (Fig. 2b and c), presumably because the
In the absence of chitosan, the mean particle diameter of electrostatic repulsion between the droplets was insufficient
the emulsion droplets was , 0.3 mm and no creaming was to prevent flocculation. At relatively high chitosan
observed at 24 h storage, which indicated that SDS was concentrations (. 0.02 wt%), where the droplets had
effective at generating small emulsion droplets during relatively high positive charges (, þ 60 mV), the
homogenization. At chitosan concentrations . 0.012 wt% measured mean particle diameters were similar to those of
there was a large increase in mean particle diameter and a the primary emulsion (, 0.3 mm). In these systems, the
decrease in turbidity, which was attributed to extensive electrical charge on the droplets must have been large
droplet aggregation and creaming (Fig. 2b and c). enough to prevent them from coming back into close
The maximum amount of droplet aggregation occurred at contact after the flocs had been disrupted by sonication.
a chitosan concentration of 0.016 wt%. In fact, droplet These experiments showed that emulsions with non-aggre-
aggregation was so extensive around this chitosan gated droplets could be produced by applying sonication to
concentration that we were unable to make reliable particle secondary emulsions containing sufficiently high levels of
size or z-potential measurements, either because the chitosan. A chitosan concentration of 0.024 wt% was
aggregates were too large or because their concentration therefore chosen to produce the secondary emulsions used
was too small to give a scattering pattern that was in the remaining experiments.
sufficiently different from that of the background. Individual
particles within these emulsions could be distinctly 3.3. Influence of pectin concentration on droplet
observed by eye, which suggested that they were at least characteristics of tertiary emulsions
100 mm in diameter. One might have expected that at
sufficiently high chitosan concentrations the emulsions The purpose of these experiments was to determine the
would have been stable to droplet aggregation because of optimum amount of pectin required to produce stable
the large electrostatic repulsion between the highly tertiary emulsions. The z-potential, particle diameter and
positively charged droplets. It is likely that the droplets turbidity of tertiary emulsions (0.2 wt% corn oil, 0.2 mM
were held together by chitosan molecules that were SDS, 0.0012 wt% chitosan, 100 mM acetic acid buffer,
adsorbed to the surface of more than one droplet (Pinotti, pH 3.0) containing different pectin concentrations
Bevilacqua, & Zaritzky 1997, 1999, 2001; Pinotti & (0 –0.04 wt%) were measured 24 h after preparation (Fig. 3).
Zaritzky, 2001). This is not surprising, since charged In the absence of pectin, the net charge on the emulsion
polymers are known to induce bridging flocculation of droplets was , þ 60 mV due to the presence of cationic
oppositely charged colloidal particles (Dickinson, 1992, chitosan on the droplet surfaces. The net charge on the
2003; Friberg et al., 2004; Von Goeler & Muthukumar, droplets changed from positive to negative as the pectin
1994). During the initial stages of mixing of the chitosan concentration in the emulsions was increased (Fig. 3a).
solution with the primary emulsion, there would have been Charge neutralization occurred when the pectin
many droplets present with completely negative surfaces; concentration was , 0.010 wt%. The negative charge on
hence, it is likely that chitosan molecules could adsorb to the the droplets reached a constant value when the pectin
surface of two or more of these droplets simultaneously. concentration exceeded , 0.020 wt%. These measurements
A similar phenomenon has been observed in previous indicated that negatively charged pectin adsorbed to the
studies (Von Goeler & Muthukumar, 1994). surface of the positively charged SDS – chitosan stabilized
In previous studies we found that sonication could be emulsion droplets. The negatively charged pectin molecules
used to effectively breakdown flocs in secondary emulsions continued to adsorb to the surfaces of the emulsion droplets,
(Ogawa et al., 2003a,b). We therefore examined the even though the droplets had no net charge or were
possibility of using the same method for producing negatively charged, up to a certain saturation level (Fig. 3a).
secondary emulsions with improved stability to droplet The ability of charged polyelectrolyte to adsorb to the
flocculation in this study. Emulsion samples (, 10 g) were surface of oppositely charged colloidal particles and cause
sonicated for 30 s at a frequency of 20 kHz, an amplitude of charge reversal is well established in the literature
40%, and a duty cycle of 0.5 s (Model 500, Sonic (Chodanowski & Stoll, 2001; Dobrynin, 2001; Jonsson &
Dismembrator, Fisher Scientific, Pittsburgh, PA), and then Linse, 2001; Kong & Muthukumar, 1998; Netz & Joanny,
the z-potential mean, mean particle diameter and turbidity 1999; Von Goeler & Muthukumar, 1994).
T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220 215

Fig. 3. (a) Dependence of particle electrical charge (z-potential) on pectin concentration for tertiary emulsions (0.2 wt% corn oil, 0.2 mM SDS, 0.0012 wt%
chitosan, 100 mM acetic acid buffer, pH 3.0). (b). Dependence of mean particle diameter ðd32 Þ on pectin concentration for tertiary emulsions (0.2 wt% corn oil,
0.2 mM SDS, 0.0012 wt% chitosan, 100 mM acetic acid buffer, pH 3.0). (c) Dependence of emulsion creaming stability on pectin concentration for tertiary
emulsions (0.2 wt% corn oil, 0.2 mM SDS, 0.0012 wt% chitosan, 100 mM acetic acid buffer, pH 3.0). The creaming stability was determined as turbidity
(at 600 nm) measured at 30% of emulsion height after 24 h storage: creaming instability is indicated by a low turbidity. The photograph shown above highlights
the instability of the emulsions to creaming at intermediate pectin concentrations (taken after sonication).

There was a significant increase (from d32 < 0:3 to thermal processing on the mean particle diameter and
1.6 mm) in the mean particle diameter and decrease in the creaming stability of primary, secondary and tertiary
turbidity (from . 0.7 to , 0.1 cm21) of the tertiary emulsions emulsions was measured after thermal processing.
when the pectin concentration was increased from 0 to There was no significant change in the mean particle
0.006 wt% (Fig. 3b and c), which corresponded to the region diameter (d32 ¼ 0:27 ^ 0.03 mm) and turbidity (p # 5%)
where charge neutralization occurred (Fig. 3a). Nevertheless of the primary (t ¼ 20:06 ^ 0.01 cm 21), secondary
at higher pectin concentration (0.02 – 0.04 wt%) (t ¼ 0:12 ^ 0.01 cm21) and tertiary emulsions (t ¼ 0:16 ^
the emulsions contained highly negatively charged particles 0.01 cm21) upon heat treatment, which indicated that they
and were stable to droplet aggregation, even in the absence of were all stable to thermal processing in the temperature
sonication. In the subsequent experiments we therefore used range used in this study.
a pectin concentration of 0.04 wt% to prepare the tertiary
emulsions. 3.5. Influence of freeze –thaw cycling on properties
of primary, secondary and tertiary emulsions
3.4. Influence of thermal processing on properties
of primary, secondary and tertiary emulsions The purpose of these experiments was to examine the
freeze –thaw stability of primary, secondary and tertiary
The purpose of these experiments was to examine the emulsions. Emulsions were frozen at 2 20 8C for 22 h, then
influence of thermal processing on the stability of SDS thawed at 30 8C for 2 h in a water bath. The influence of
(primary emulsion), SDS – chitosan (secondary emulsion) freeze –thaw cycling (up to 6 times) on the mean particle
and SDS – chitosan – pectin (tertiary emulsion) coated diameter and creaming stability of diluted primary,
droplets. Primary, secondary and tertiary emulsions with secondary and tertiary emulsions at pH 3.0 was measured
the same oil concentration (0.2 wt%) were held at (Fig. 4). The primary and secondary emulsions were highly
temperatures ranging from 30 to 90 8C, cooled to room unstable to freeze – thaw cycling, there being an appreciable
temperature, and then stored for 24 h. The influence of increase in mean particle diameter (Fig. 4a) and rapid
216 T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220

Fig. 4. (a) Dependence of mean particle diameter of primary, secondary and tertiary emulsions on number of freeze–thaw cycles. (b) Dependence of emulsion
creaming stability of primary, secondary and tertiary emulsions on number of freeze– thaw cycles. The creaming stability was determined as turbidity
(at 600 nm) measured at 30% of emulsion height after 24 h storage: creaming instability is indicated by a low turbidity. A photograph of the primary, secondary
and tertiary emulsion after 1 freeze–thaw cycle is shown above.

creaming (Fig. 4b) after the first cycle, suggesting extensive of freeze –thaw cycling on emulsion stability. First, water
droplet aggregation. Nevertheless, the particle size data on crystallization occurred when the emulsions were placed in
the primary and secondary emulsions after freezing and the freezer. As more and more water crystallized the droplets
thawing should be treated with caution because they were would have been forced closer together (Saito et al., 1999)
either highly coalesced (exhibiting oiling off) or highly and there may not have been sufficient free water present to
flocculated, respectively, thereby making reliable light fully hydrate the droplet surfaces (Ausborn et al., 1994;
scattering measurements difficult to make. On the other Komatsu, Okada, & Handa 1997; Strauss & Hauser, 1986),
hand, the mean particle diameter of the particles in the thus favoring droplet – droplet interactions. Second, ice
tertiary emulsions remained stable after 6 freeze – thaw crystallization leads to an increase in the ionic strength of
cycles (, 0.3 ^ 0.05 mm), and there was no evidence of any freeze-concentrated non-frozen aqueous phase
creaming (the turbidity remained fairly constant). surrounding the emulsion droplets (Komatsu et al., 1997),
A number of physicochemical phenomena that occur which may have promoted droplet –droplet interactions.
during frozen storage may account for the observed influence Third, it is possible that ice crystals formed during freezing
T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220 217

may have penetrated into the oil droplets and disrupted their high salt concentrations than droplets coated with SDS or
interfacial membranes, thus making them more prone to SDS –chitosan membranes, which may have been important
coalescence. Fourth, cooling may have caused some of the fat if the ionic strength of the freeze-concentrated solution
in the emulsion droplets to crystallize, which may have surrounding the droplets increased (see Section 3.6).
promoted partial coalescence due to penetration of a fat
crystal from one droplet through the membrane of another 3.6. Influence of ionic strength on properties of primary,
droplet (Harada & Yokomizo, 2000; Vanapalli, Palanuwech, secondary and tertiary emulsions
& Coupland 2002).
There are a number of possible reasons that the tertiary The purpose of these experiments was to examine the
emulsions were more stable to freeze –thaw cycling than influence of ionic strength on the stability of primary,
the primary and secondary emulsions. First, the interfacial secondary and tertiary emulsions. The influence of NaCl
layer in the tertiary emulsions is thicker than that in the concentration (0 – 500 mM) on the z-potential, mean particle
primary and the secondary emulsions, so there will be a diameter and turbidity of diluted primary, secondary and
greater short-range steric repulsion between the droplets tertiary emulsions at pH 3.0 was measured (Fig. 5).
that prevents them from coming close enough to The z-potential of the SDS-stabilized droplets in the primary
coalesce (Hunter, 1986; Israelachvili, 1992). Second, it emulsions was negative at all ionic strengths, but the
may be more difficult for fat or ice crystals to rupture the magnitude of the z-potential decreased as the NaCl
thicker SDS – chitosan– pectin membranes than the thinner concentration was increased which can be attributed to
SDS or SDS – chitosan membranes, thereby making the electrostatic screening (Hunter, 1986). The z-potential of
droplets in the tertiary emulsion more stable to the chitosan –SDS stabilized droplets in the secondary
droplet coalescence or partial coalescence (Goff, 2000; emulsions was positive at all ionic strengths but the
McClements, 2001). Third, droplets coated with SDS –chi- magnitude of the z-potential decreased as the NaCl
tosan –pectin membranes are more stable to aggregation at concentration was increased, which can also be attributed

Fig. 5. (a) Dependence of particle electrical charge (z-potential) on NaCl concentration for primary, secondary and tertiary emulsions at pH 3. (b) Dependence
of mean particle diameter ðd32 Þ on NaCl concentration for primary, secondary and tertiary emulsions at pH 3. (c) Dependence of emulsion creaming stability on
NaCl concentration for primary, secondary and tertiary emulsions at pH 3. The creaming stability was determined as turbidity (at 600 nm) measured at 30% of
emulsion height after 24 h storage: creaming instability is indicated by a low turbidity. The photograph shown above highlights the instability of the primary
emulsions to creaming at high NaCl concentrations.
218 T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220

to electrostatic screening effects (Hunter, 1986). It is also


possible that some of the decrease of charge on the emulsion
droplets in the secondary emulsions was due to desorption
of chitosan from the emulsion droplet surfaces when the
ionic strength increased. This could have occurred because
the electrostatic attraction between the negatively charged
SDS molecules and the positively charged chitosan
molecules was weakened at higher ionic strengths,
thus allowing desorption of some of the cationic
biopolymer. The z-potential of the pectin – chitosan– SDS
stabilized droplets in the tertiary emulsion was negative at
all ionic strengths but the magnitude of the z-potential did
not change appreciatively as the NaCl concentration was
increased (Fig. 5a), which suggested that pectin did not
desorb from the emulsion droplet surfaces at higher ionic
strength. In addition, one would have expected a decrease in
z-potential due to electrostatic screening effects, which was
not observed. It is possible that the thickness of this
adsorbed interfacial layer changed with ionic
strength, which kept the magnitude of the surface potential
relatively constant.
The secondary and tertiary emulsions were relatively
stable to droplet aggregation at all ionic strengths
(0 – 500 mM NaCl), as demonstrated by the relative
independence of turbidity on NaCl concentration (Fig. 5c).
On the other hand the primary emulsion became strongly
aggregated at high ionic strength (500 mM NaCl), as shown
by the large decrease in turbidity (Fig. 5c). In the case of the
primary emulsion, droplet aggregation was probably caused
by the reduction of repulsive interactions between the
droplets due to electrostatic screening, but it may also have
been due to the ability of salt to change the optimum
curvature of the SDS (Hunter, 1986; Israelachvili, 1992).
The secondary and tertiary emulsions were probably
stable to droplets aggregation at high salt because the
relatively thick interfacial membranes provided good steric
stabilization.

3.7. Influence of pH on stability of primary, secondary Fig. 6. (a) Dependence of particle electrical charge (z-potential) on pH
and tertiary emulsions concentration for primary, secondary and tertiary emulsions. (b) Depen-
dence of mean particle diameter ðd32 Þ on pH for primary, secondary and
The purpose of these experiments was to examine the tertiary emulsions. (c) Dependence of emulsion creaming stability on pH
for primary, secondary and tertiary emulsions after 10 days storage. The
influence of pH on the stability of primary, secondary and creaming stability was determined as turbidity (at 600 nm) measured at
tertiary emulsions. The influence of pH on the z-potential, 30% of emulsion height after 24 h storage: creaming instability is indicated
mean particle diameter, and turbidity of diluted primary, by a low turbidity.
secondary and tertiary emulsions at pH 3.0 – 8.0 was
measured after they had been stored at room temperature (pH 7 and 8). The reversal of charge on the emulsion
for 1 day and 10 days after preparation (Fig. 6). droplets suggests that the chitosan was desorbed from the
The z-potential of the SDS stabilized droplets in primary emulsion droplet surfaces when the pH was increased above
emulsions was negative at all pH values, but the magnitude 6. The cationic groups (– NHþ3 ) on chitosan typically have a
of the z-potential decreased when the pH was increased pKa value around pH 6.3 –7; hence, the chitosan loses its
from 3 to 8, probably due to electrostatic screening effects charge around this pH (Ogawa et al., 2004). The z-potential
associated with addition of the alkali used to adjust pH of the pectin – chitosan– SDS stabilized droplets in the
(Hunter, 1986). The z-potential of the chitosan – SDS tertiary emulsion was negative at all pH, which suggested
stabilized droplets in the secondary emulsions was positive that the chitosan layer did not desorb from the surface of
at low pH (, pH 6), but became negative at higher values the emulsion droplets at higher pH values even though it
T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220 219

did in the secondary emulsions. This may have been because droplet aggregation and creaming at high temperatures
the pKa value of the positively charged groups on the (90 8C), several freeze – thaw cycles, high salt
chitosan molecules was increased appreciably when the concentrations (NaCl 500 mM), and a wide range of pH
chitosan was sandwiched between two negatively charged values (pH 3 –8). The improved stability of the tertiary
biopolymer layers, as has been reported for other emulsions can be attributed to the ability of multilayered
polyelectrolytes (Burke & Barrett, 2003a,b). The increase interfacial membranes to either increase the repulsive
in the magnitude of the negative charge with increasing pH, colloidal interactions between the droplets (e.g. electrostatic
was probably because of the increase in the negative charge and steric) and/or to increase the resistance of the membrane
on the pectin molecules (pKa , 4.5) and the decrease in to rupture. The tertiary emulsions may have a number of
the positive charge on the chitosan molecules (pKa , 6.5) as potential applications in the food industry for specialized
the pH increased. applications. For example, the thick interfacial membranes
The droplets in the primary, secondary and tertiary may be useful in protecting lipid components, for controlled
emulsions were stable to droplet aggregation at pH 3 –5 flavor release or for improving the stability of emulsions to
when stored for 10 days. But aggregation occurred in environmental stresses, such as freezing and heating.
primary and secondary emulsion after 10 days storage at pH Finally, we should point out some of the current
6 –8 as demonstrated by the rapid decrease in creaming limitations of the layer-by-layer deposition method for
stability in these emulsions (Fig. 6b). This was probably application in the food industry. The final emulsions used in
because the decrease in droplet charge (Fig. 6a) reduced the this study had relatively low droplet concentrations
electrostatic repulsion between the droplets, which led to (0.2 wt%), which may be suitable for certain applications
extensive droplet flocculation. The tertiary emulsion was in the food industry (e.g. beverages), but not for others
stable to creaming and droplet aggregation over the whole (e.g. dressings, sauces, soups). One of the major problems
pH range because of the relatively strong electrostatic and with preparing more concentrated emulsions is the
steric repulsion associated with the relatively thick and tendency for the droplets to flocculate due to a bridging
electrically charged three-layer interfacial membrane. mechanism. We are currently investigating methods of
In addition, the creaming stability may have been improved optimizing the preparation procedures so that more
because the multiple layers increased the overall density of concentrated emulsions can be prepared. Another potential
the droplets, thereby reducing the density contrast between limitation of this technology is the number of additional
the dispersed and continuous phases and decreasing the steps required to prepare the multilayered emulsions
driving force for gravitational separation (Friberg et al., compared to conventional emulsions, which will increase
2004; McClements, 1999; Stauffer, 1999). production costs and therefore limit its application to certain
specialized products.

4. Conclusions
Acknowledgements
This study has shown that stable emulsions containing
tri-layered lipid droplets can be prepared using a simple and
This material is based upon work supported by the
cost-effective method. Initially, a primary emulsion
National Research Initiative, Competitive Grants Program,
containing small droplets was produced by homogenization
United States Department of Agriculture, and the CREES,
of oil, water, and an anionic emulsifier (SDS at pH 3.0).
IFAFS Program, United States Department of Agriculture
A secondary emulsion containing cationic lipid droplets
(Award Number 2002-35503-12296). We also thank the
coated with a SDS –chitosan membrane was then produced
Tokyo University of Marine Science and Technology
by mixing a cationic polysaccharide (chitosan at pH 3) with
(Japan) for financial support of Tomoko Aoki.
the primary emulsion and agitating to disrupt any flocs
formed. A tertiary emulsion containing anionic droplets
coated with a SDS – chitosan – pectin membrane was then
produced by mixing an anionic polysaccharide (pectin at pH References
3) with the secondary emulsion, without applying mechan-
Ausborn, M., Schreier, H., Brezesinski, G., Fabian, H., Meyer, H. W., &
ical agitation.
Nuhn, P. (1994). The protective effect of free and membrane-bound
This study also compared the stability of SDS-coated cryoprotectants during freezing and freeze-drying of liposomes.
(primary), SDS – chitosan-coated (secondary) and Journal of Controlled Release, 30, 105 –116.
SDS – chitosan – pectin-coated (tertiary) oil droplets to Burke, S. E., & Barrett, C. J. (2003a). pH-responsive properties of
droplet aggregation induced by thermal processing, multilayered poly(L -lysine)/hyaluronic acid surfaces. Biomacromole-
cules, 4, 1773–1783.
freeze – thaw cycling, NaCl and pH. Our results clearly
Burke, S. E., & Barrett, C. J. (2003b). Acid–base equilibria of weak
show that emulsion stability could be dramatically polyelectrolytes in multilayer thin films. Langmuir, 19, 3297–3303.
improved by coating the droplets with two or more layers. Calvo, P., Remunan-Lopez, C., Vila-Jato, J. L., & Alonso, M. J. (1997).
The droplets in the tertiary emulsions had good stability to Development of positively charged colloidal drug carriers:
220 T. Aoki et al. / Food Hydrocolloids 19 (2005) 209–220

Chitosan-copolyester nanocapsules and submiron-emulsions. Moreau, L., Kim, H.-J., Decker, E. A., & McClements, D. J. (2003).
Colloid and Polymer Science, 275, 46–53. Production and characterization of oil-in-water emulsions containing
Caruso, F. (2001). Generation of complex colloids by polyelectrolyte- droplets stabilized by b-lactoglobulin–pectin membranes. Journal of
assisted electrostatic self-assembly. Australian Journal of Chemistry, Agricultural and Food Chemistry, 51, 6612– 6617.
54, 349 –353. Netz, R. R., & Joanny, J.-F. (1999). Complexation between a semiflexible
Caruso, F., & Schuler, C. (2000). Enzyme multilayers on colloid particles: polyelectrolyte and an oppositely charged sphere. Macromolecules, 32,
Assembly, stability, and enzymatic activity. Langmuir, 16, 9595–9603. 9026–9040.
Charlambous, G., & Doxastakis, G. (1989). Food emulsifiers: Chemistry, Ogawa, S., Decker, E. A., & McClements, D. J. (2003a). Production and
technology, functional properties and applications. Amsterdam: characterization of O/W emulsions containing cationic droplets
Elsevier. stabilized by lecithin– chitosan membranes. Journal of Agricultural
Chodanowski, P., & Stoll, S. (2001). Polyelectrolyte adsorption on charged and Food Chemistry, 51, 2806–2812.
particles: Ionic concentration and particle size effects—A Monte Carlo Ogawa, S., Decker, E. A., & McClements, D. J. (2003b). Influence of
approach. Journal of Chemical Physics, 115, 4951–4960. environmental conditions on stability of O/W emulsions containing
Dickinson, E. (1992). An introduction to food colloids. Oxford: droplets stabilized by lecithin – chitosan membranes. Journal of
Oxford Science Publishers. Agricultural and Food Chemistry, 51, 5522– 5527.
Dickinson, E. (2003). Hydrocolloids at interfaces and the influence on the Ogawa, S., Decker, E. A., & McClements, D. J. (2004). Production and
properties of dispersed systems. Food Hydrocolloids, 17(1), 25–39. characterization of O/W emulsions containing droplets stabilized by
Dickinson, E., & James, J. D. (2000). Influence of high-pressure treatment lecithin–chitosan–pectin mutilayered membranes. Journal of Agricul-
on b-lactoglobulin–pectin associations in emulsions and gels. Food tural and Food Chemistry, Vol. 52. pp. 3595–3600.
Hydrocolloids, 14, 365–376. Pinotti, A., Bevilacqua, A., & Zaritzky, N. (1997). Optimization of the
Dobrynin, A. V. (2001). Effect of solvent quality on polyelectrolyte flocculation stage in a model system of a food emulsion waste using
adsorption at an oppositely charged surface. Journal of Chemical chitosan as polyelectrolyte. Journal of Food Engineering, 32, 69–81.
Physics, 115, 8145–8153. Pinotti, A., Bevilacqua, A., & Zaritzky, N. (1999). Treatment of anionic
Faldt, P., Bergenstahl, B., & Claesson, P. M. (1993). Stabilization by emulsion systems using chitosan, polyacrylamide, and aluminum
chitosan of soybean oil emulsions coated with phospholipid and sulfate. Scanning, 21, 354 –358.
glycholic acid. Colloids and Surfaces A, 71, 187–195. Pinotti, A., Bevilacqua, A., & Zaritzky, N. (2001). Comparison of the
Friberg, S., Larsson, K., & Sjoblom, J. (2004). Food emulsions (4th ed.). performance of chitosan and a cationic polyacrylamide as flocculants of
New York: Marcel Dekker. emulsion systems. Journal of Surfactants and Detergents, 4, 57–63.
Garti, N., & Reichman, D. (1993). Hydrocolloids as food emulsifiers and Pinotti, A., & Zaritzky, N. (2001). Effect of aluminum sulfate and cationic
stabilizers. Food Microstructure, 12, 411 –426. polyelectrolytes on the destabilization of emulsified wastes. Waste
Goff, H. D. (2000). Controlling ice-cream structure by examining fat: Management, 21, 535–542.
Protein interactions. Australian Journal of Dairy Technology, 55, Saito, H., Kawagishi, A., Tanaka, M., Tanimoto, T., Okada, S., Komatsu,
78–81. H., & Handa, T. (1999). Coalescence of lipid emulsions in floating and
Harada, T., & Yokomizo, K. (2000). Demulsification of oil-in-water freeze-thawing processes: Examination of the coalescence transition
emulsion under freezing conditions: Effect of crystal structure state theory. Journal of Colloid and Interface Science, 219, 129 –134.
modifier. Journal of the American Oil Chemical Society, 77, Schmitt, C., Sanchez, C., Desobry-Banon, S., & Hardy, J. (1998). Structure
859– 863. and technofunctional properties of protein–polysaccharide complexes.
Hunter, R. J. (1986) (Vol. 1). Foundations of colloid science, Oxford: Critical Reviews in Food Science and Nutrition, 38, 689–753.
Oxford University Press. Shahidi, F., Arachchi, J. K. V., & Jeon, Y. J. (1999). Food applications of
Israelachvili, J. N. (1992). Intermolecular and surface forces. London: chitin and chitosans. Trends in Food Science and Technology, 10,
Academic Press. 37 –51.
Jonsson, M., & Linse, P. (2001). Polyelectrolyte –macroion complexation: Shi, X. Y., & Caruso, F. (2001). Release behavior of thin-walled
Effect of linear charge density and macroion charge. Journal of microcapsules composed of polyelectrolyte multilayers. Langmuir,
Chemical Physics, 115, 3406–3418. 17, 2036–2042.
Komatsu, H., Okada, S., & Handa, T. (1997). Suppressive effects of salts Skjak-Braek, G., Anthonsen, T., & Sandford, P. (1989). Chitin and
on droplet coalescence in a commercially available fat emulsion chitosan. London: Elsevier.
during freezing for storage. Journal of Pharmaceutical Science, 86, Stauffer, C. E. (1999). Emulsifiers. St Paul, MN: Eagen Press.
497– 502. Strauss, G., & Hauser, H. (1986). Stabilization of lipid bilayer vesicles by
Kong, K. Y., & Muthukumar, M. (1998). Monte Carlo study of adsorption sucrose during freezing. Proceedings of the National Academy of
of a polyelectrolyte onto charged spheres. Journal of Chemical Physics, Science USA, 83, 2422–2426.
109, 1522–1527. Vanapalli, S. A., Palanuwech, J., & Coupland, J. N. (2002). Stability of
Krog, N. J., & Sparso, F. V. (2004). Food emulsifiers: Their chemical and emulsions to dispersed phase crystallization: Effect of oil type,
physical properties. In S. Friberg, K. Larsson, & J. Sjoblom (Eds.), dispersed phase volume fraction, and cooling rate. Colloids and
Food emulsions (4th ed.). New York: Marcel Dekker, chapter 2. Surfaces A, 204, 227–237.
Magdassi, S., Bach, U., & Mumcuoglu, K. Y. (1997). Formation of Von Goeler, F., & Muthukumar, M. (1994). Adsorption of polyelectrolytes
positively charged microcapsules based on chitosan–lecithin inter- onto curved spheres. Journal of Chemical Physics, 100, 7796–7803.
actions. Journal of Microencapsulation, 14, 189–195. Walstra, P. (1993). Principles of emulsion formation. Chemical Engineer-
McClements, D. J. (1999). Food emulsions: Principles, practice and ing and Science, 48, 333–350.
techniques. Boca Raton, FL: CRC Press. Walstra, P (2003). Physical chemistry of foods. New York: Marcel Dekker.
McClements, D. J. (2001). Estimation of steric exclusion and differential Williams, P. A, & Phillips, G. O. (2003). The use of hydrocolloids to
interaction contributions to protein transfer free energies in aqueous improve food texture. In B. M. McKenna (Ed.), Texture in foods,
cosolvent solutions. Food Hydrocolloids, 15, 355 –363. volume 1: Semi-solid foods, Boca Raton, FL: CRC Press, Chapter 11.

You might also like